Mining Platoon Patterns from Traffic Videos

TL;DR

This work addresses co-movement pattern mining from video data by relaxing the requirement of consecutive cameras along a common route, introducing VPlatoon with parameters $m$, $k$, $d$, and $\epsilon$. It introduces MaxGrowth, a cluster-sequence enumeration framework that directly generates maximal relaxed patterns without a costly candidate verification step, and two pruning rules (Root and Dependency) to prune non-maximal patterns. Empirical results on real and semi-synthetic datasets show that MaxGrowth is up to two orders of magnitude faster than the baseline FRB method and yields higher-pattern quality under imperfect trajectory recovery. The approach improves robustness to occlusions and ID switches, offering practical benefits for smart city surveillance and traffic management. Overall, the paper demonstrates that a cluster-based enumeration strategy with targeted pruning can efficiently mine meaningful, maximal co-movement patterns from noisy video-derived trajectories.

Abstract

Discovering co-movement patterns from urban-scale video data sources has emerged as an attractive topic. This task aims to identify groups of objects that travel together along a common route, which offers effective support for government agencies in enhancing smart city management. However, the previous work has made a strong assumption on the accuracy of recovered trajectories from videos and their co-movement pattern definition requires the group of objects to appear across consecutive cameras along the common route. In practice, this often leads to missing patterns if a vehicle is not correctly identified from a certain camera due to object occlusion or vehicle mis-matching. To address this challenge, we propose a relaxed definition of co-movement patterns from video data, which removes the consecutiveness requirement in the common route and accommodates a certain number of missing captured cameras for objects within the group. Moreover, a novel enumeration framework called MaxGrowth is developed to efficiently retrieve the relaxed patterns. Unlike previous filter-and-refine frameworks comprising both candidate enumeration and subsequent candidate verification procedures, MaxGrowth incurs no verification cost for the candidate patterns. It treats the co-movement pattern as an equivalent sequence of clusters, enumerating candidates with increasing sequence length while avoiding the generation of any false positives. Additionally, we also propose two effective pruning rules to efficiently filter the non-maximal patterns. Extensive experiments are conducted to validate the efficiency of MaxGrowth and the quality of its generated co-movement patterns. Our MaxGrowth runs up to two orders of magnitude faster than the baseline algorithm. It also demonstrates high accuracy in real video dataset when the trajectory recovery algorithm is not perfect.

Figures (16)

• Figure 1: An illustrative example for relaxed co-movement pattern mining from real video data.
• Figure 2: Travel paths for four objects $\{o_1,o_2,o_3,o_4\}$.
• Figure 3: Example clusters under parameters $m = 2$ and $\epsilon = 6$.
• Figure 4: The search tree of MaxGrowth under parameters $m = 2$, $k = 2$, $d = 1$, and $\epsilon$ = 6.
• Figure 5: Varying $m$.

Theorems & Definitions (28)

• Definition 1
• Example 1
• Definition 2
• Example 2
• Definition 3
• Example 3
• Definition 4
• Example 4
• Definition 5
• Example 5